AI-Assisted Zero Trust Architecture for Continuous Risk Assessment of Programmable Logic Controllers in Food Processing Infrastructure

Food processing infrastructure relies heavily on programmable logic controllers (PLCs) to manage physical processes that support safety and continuous operation. As these environments become increasingly interconnected through the integration of information technology and operational technology, cyber risks continue to grow. Traditional perimeter-based security approaches and static Zero Trust implementations often struggle to address continuously changing controller behavior in operational environments. In addition, many PLCs lack built-in security capabilities that provide the device-level visibility needed for ongoing trust evaluation. This paper presents a conceptual architecture that combines artificial intelligence (AI) with Zero Trust principles to support continuous monitoring and risk assessment of PLCs in food processing environments. The proposed approach uses passive behavioral observation and network-level monitoring to track controller activity over time and enable adaptive trust decisions based on observed risk. By improving visibility without requiring changes to controller hardware or control logic, the architecture helps preserve safety and availability. Future work will focus on validating the proposed architecture in simulated operational technology environments and assessing its impact on detection accuracy, safety, and system performance.